Am J Cancer Res 2015;5(9):2643-2659 www.ajcr.us /ISSN:2156-6976/ajcr0008500

Original Article LIN28B suppresses microRNA let-7b expression to promote CD44+/LIN28B+ stem cell proliferation and invasion

Yebo Shao1*, Lei Zhang1*, Lei Cui2*, Wenhui Lou1, Dansong Wang1, Weiqi Lu1, Dayong Jin1, Te Liu3

1Department of General Surgery, Zhongshan Hospital, Fudan University, Shanghai 200032, China; 2Department of General Surgery, Jiangsu University Affiliated Hospital, Zhengjiang 212000, China; 3Shanghai Geriatric Institute of Chinese Medicine, Longhua Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai 200031, China. *Equal contributors. Received March 27, 2015; Accepted July 5, 2015; Epub August 15, 2015; Published September 1, 2015

Abstract: Although the highly proliferative, migratory, and multi-drug resistant phenotype of human pancreatic can- cer stem cells (PCSCs) is well characterized, knowledge of their biological mechanisms is limited. We used CD44 and LIN28B as markers to screen, isolate, and enrich CSCs from human primary pancreatic cancer. Using flow cytometry, we identified a human primary pancreatic (PCC) subpopulation expressing high levels of both CD44 and LIN28B. CD44+/LIN28B+ PCSCs expressed high levels of stemness marker and possessed higher migratory and invasive ability than CD44-/LIN28B- PCCs. CD44+/LIN28B+ PCSCs were more resistant to growth inhibition induced by the chemotherapeutic drugs cisplatin and gemcitabine hydrochloride, and readily established tumors in vivo in a relatively short time. Moreover, microarray analysis revealed significant differences between the cDNA expression patterns of CD44+/LIN28B+ PCSCs and CD44-/LIN28B- PCCs. Following siRNA interference of endogenous Lin28b expression in CD44+/LIN28B+ PCSCs, not only was their proliferation decreased, there was also cell cycle arrest due to suppression of cyclin D1 expression following the stimulation of miRNA let-7b expression. In conclusion, CD44+/LIN28B+ cells, which possess CSC characteristics, can be reliably sorted from human primary PCCs and represent a valuable model for studying cancer cell physiology and multi-drug resistance.

Keywords: Pancreatic ductal , cancer stem cells, CD44, Lin28B

Introduction properties; furthermore, the capacity for tumor formation and growth resides exclusively in a Pancreatic ductal adenocarcinoma (PDAC) is a small proportion of tumor cells, termed cancer highly lethal malignancy that is usually diag- stem cells (CSCs) [7-11]. Pancreatic CSCs nosed at a late stage, at which optimal thera- (PCSCs) were first characterized by Li et al. 6[ ] peutic options have been skipped [1]. It is one and were shown to be not only highly tumori- of the most chemoresistant tumors; the surviv- genic, but also possessed the ability to self- al rate is < 5% [2]. PDAC is not only notorious renew and produce differentiated progeny that for being difficult to diagnose at an early stage reflected the heterogeneity of the patient’s pri- and its poor recurrence-free prognosis, but mary tumor [3]. Moreover, an increasing num- also the lack of effective treatment thereof and ber of studies have reported human pancreatic limited knowledge of its biological characteris- cancer cell subpopulations, such as CD31+/ tics [3, 4]. Thus, there is an urgent need for bet- CD45+ [12]; Hoechst 33342-/CD133+/ALDH1+ ter understanding of the cellular/molecular [1]; ESA+/CD44+ [3]; and CD24+/CD44+ [4], properties associated with PDAC to explore which express CSC-associated characteristics novel venues of diagnosis and treatment of this and stem cell markers [5]. There are several disease [1, 5, 6]. Recent evidence suggests prominent CSC characteristics: they (a) self- that tumors consist of heterogeneous cell pop- renew and are highly clonogenic, (b) differenti- ulations, which possess different biological ate in vitro to form organized spheroids in sus- LIN28B promotion of PCSC proliferation and invasion pension, (c) express multipotency and tissue- We demonstrated a CD44+/LIN28B+ PCSC specific differentiation markers, (d) generate subpopulation that proliferates rapidly and tumors in vivo through self-renewal mecha- exhibits multi-drug resistance, high invasion nisms, (e) undergo in vivo differentiation to pro- ability, and adherin. Therefore, CD44+/LIN28B+ duce a disease similar to that in the patient PCSCs represent a potentially powerful in vitro [13]. The observation that stem cells and some model for studying cancer cell metastasis, inva- CSCs share the common defining features of sion, and self-renewal and for assessing the incompletely differentiated state and self- effectiveness of novel therapeutics for PDAC. renewal capacity led to the CSC hypothesis as a possible mechanism for total tumor growth Materials and methods as the result of the proliferation of a small sub- Isolation CD44 and LIN28B phenotype cells by population of cells [9-11, 14]. magnetic activated cell sorting system Lin28, which is an RNA-binding , regu- lates cell growth and differentiation [15]. CD44+ and LIN28B+ subpopulation cells were Developmental timing in Caenorhabditis ele- isolated from primary cancer cells from pancre- gans is regulated by a heterochronic gene path- atic cancer tissues using 4 μl of the primary monoclonal antibodies (rabbit anti-human way. The heterochronic gene lin28 is a key LIN28B-FITC, rabbit anti-human CD44-PE, regulator early in the pathway [16]. lin28 eBioscience) stored at 4°C in PBS for 30 min in encodes an approximately 25-kDa protein with a volume of 1 ml as previously described [7, two RNA-binding motifs: a so-called “cold shock 21]. After reaction, the cells were washed twice domain” (CSD) and a pair of retroviral-type in PBS, and were put the secondary monoclo- CCHC zinc fingers; it is the only known animal nal antibodies (Goat anti-rabbit coupled to protein with this motif pairing. The CSD is a magnetic microbeads, Miltenyi Biotec, Auburn, β-barrel structure that binds single-stranded CA), incubated at 10°C in PBS for 15 min and nucleic acids [16]. Lin28 inhibits the biogene- then washed twice in PBS. Single cells were sis of a group of microRNAs (miRNAs), among plated at 1000 cells/ml in DMEM: F12 which are the let-7 family miRNAs shown to par- (HyClone), supplemented with 10 ng/ml basic ticipate in regulation of the expression of genes fibroblast growth factor (bFGF), 10 ng/ml epi- involved in cell growth and differentiation [17]. dermal growth factor (EGF), 5 μg/ml insulin and The mechanism underlying selective let-7 inhi- 0.5% bovine serum albumin (BSA) (all from bition by Lin28 has been studied extensively. Sigma-Aldrich). All CD44+/LIN28B+ cells were The common theme is that Lin28 binds to the cultured in above conditions as non-adherent terminal loop region of pri/pre-let-7 and blocks spherical clusters which were called PCSCs, their processing [15]. The miRNAs are small and CD44-/LIN28B- cells which were cultured RNA molecules (21-23 nucleotides) that act as under general conditions as adherent clusters, negative regulators of gene expression either was called PCCs. All Cells had been cultured on by blocking mRNA translation into protein or the same conditions until passage 4th before through RNA interference [18-21]. Previous making ulterior experiments. The methods studies have reported that dysregulation of were carried out in accordance with the specific miRNAs is associated with certain approved guidelines. types of cancer, and they are thought to act as either oncogenes or tumor suppressors Quantitative Real-time PCR (qRT-PCR) analysis depending on the target gene [19, 21, 22]. Furthermore, the miRNA let-7b regulates self- Total RNA from each cells was isolated using renewal of embryonic stem cells and the prolif- Trizol Reagent (Invitrogen) according to the eration and tumorigenicity of cancer cells by manufacturer’s protocol. The RNA samples inhibiting cyclin D1 (CCND1) expression were treated with Dnase I (Sigma-Aldrich), [23-25]. quantified, and reverse-transcribed into cDNA using the ReverTra Ace-α First Strand cDNA In view of the above findings, we sorted a novel Synthesis Kit (TOYOBO). qRT-PCR was conduct- CSC subpopulation overexpressing CD44 and ed using a RealPlex4 real-time PCR detection LIN28B at the cell surface (CD44+/LIN28B+) system from Eppendorf Co. LTD (Germany), from human primary pancreatic cancer tissues. with SyBR Green RealTime PCR Master MIX

2644 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

(TOYOBO) used as the detection dye. qRT-PCR DAPI (Sigma-Aldrich) at room temperature for amplification was performed over 40 cycles 30 min. Then the cells were thoroughly washed with denaturation at 95°C for 15 sec and with TBST and viewed through a fluorescence annealing at 58°C for 45 sec. Target cDNA was microscope (DMI3000; Leica, Allendale, NJ, quantified using the relative quantification USA). method. A comparative threshold cycle (Ct) was used to determine gene expression relative to a Soft agar colony formation assay control (calibrator) and steady-state mRNA lev- els are reported as an n-fold difference relative All steps were according to the previously to the calibrator. For each sample, the maker described [19]. Soft Agar Assays were con- genes Ct values were normalized using the for- structed in 6-well plates. The base layer of each mula ΔCt = Ct_markers-Ct_18sRNA. To deter- well consisted of 2 mL with final concentrations mine relative expression levels, the following of 1 × media (DMEM+10% FBS) and 0.6% low formula was used ΔΔCt = ΔCt_CSCs-ΔCt_CCs. melting point agarose. Plates were chilled at The values used to plot relative expressions of 4°C until solid. Upon this, a 1.0 ml growth agar 4 markers were calculated using the expression layer was poured, consisting of 1 × 10 cells 2-ΔΔCt. The mRNA levels were calibrated based suspended in 1 × media and 0.3% low melting on levels of 18 s RNA. The cDNA of each stem point agarose. Plates were again chilled at 4°C cell markers was amplified using primers as until the growth layer congealed. An additional previously described [21]. 1.0 ml of 1 × media without agarose was added on top of the growth layer on day 0 and again Multi-chemodrugs resistant assay on day 15 of growth. Cells were allowed to grow at 37°C for 1 month and total colonies counted. The chemodrugs (casplatin, gemcitabine hydro- Assays were repeated a total of 3 times. Results chloride) resistant assay of each cell was com- were statistically analyzed by paired T-test pletely performed as previously described [7, using the PRISM Graphpad program. 21]. Transwell migration assay Western blotting analysis All steps were according to the previously Protein extracts of each cell were resolved described [19]. Cells (2 × 105) were resuspend- by 12% SDS-PAGE and transferred on PVDF ed in 200 μl of serum-free medium and seeded (Millipore) membranes. After blocking, the on the top chamber of the 8.0 μm pore, 6.5 mm PVDF membranes were washed 4 times for 15 polycarbonate transwell filters (Corning). The min with TBST at room temperature and incu- full medium (600 μl) containing 10% FBS was bated with primary antibody (rabbit anti-human added to the bottom chamber. The cells were LIN28B, rabbit anti-human CCND1 all from allowed to migrate for 24 h at 37°C in a humidi-

Cell Signaling Technology). Following extensive fied incubator with 5% CO2. The cells attached washing, membranes were incubated with sec- to the lower surface of membrane were fixed in ondary peroxidase-linked Goat anti- rabbit IgG 4% paraformaldehyde at room temperature for (Santa Cruz) for 1 h. After washing 4 times for 30 mins and stained with 4,6-diamidino-2-phe- 15 min with TBST at room temperature once nylindole (DAPI) (C1002, Beyotime Inst Biotech, more, the immunoreactivity was visualized by China), and the number of cells on the lower enhanced chemiluminescence (ECL kit, Pierce surface of the filters was counted under the Biotechnology). microscope. A total of 5 fields were counted for each transwell filter. Immunofluorescence staining analysis In vivo xenograft experiments The cultured cells were washed 3 times with PBS and fixed with 4% paraformaldehyde About 1 × 104 cells (PCSCs or PCCs) were inoc- (Sigma-Aldrich, St. Louis, USA) for 30 min. After ulated s.c in athymic nude mice. All mice of 6-7 blocking, the cells were incubated first antibod- weeks of age were carried out at Experimental ies overnight at 4°C, and then with Cy3- Animal Center of Fudan University and Use conjugated goat anti-rabbit IgG antibody Committer approval in accordance with institu- (1:200; Abcam, Cambridge, UK) and 5 μg/ml tional guidelines.

2645 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Northern blotting ethanol, and kept in a freezer more than 48 h. Before flow cytometric analysis, The fixed cells Northern blotting was done as previously were centrifuged, washed twice with PBS, and described [20]. For all cell treatment groups, resuspended in PI staining solution (Sigma- 20 μg of good quality total RNA was analyzed Aldrich Chemical) containing 50 μL/mL PI and on a 7.5 M urea, 12% PAA denaturing gel and 250 μg/mL RNase A (Sigma-Aldrich Chemical). transferred to a Hybond N+ nylon membrane The cell suspension, which was hidden from (Amersham, Freiburg, Germany). Membranes light, were incubated for 30 min at 4°C and were cross-linked using ultraviolet light for 30 s analyzed using the FACS (FACSAria, BD at 1200 mJ/cm2 and hybridized to the let-7b Bioscience, CA, USA). A total of 20,000 events antisense Starfire probe, 5’-AACCACACAACC- were acquired for analysis using CellQuest TACTACCTCA-3’ (Sangon Biotech Co., Ltd, software. Shanghai, China) for the detection of 22-nt le7- 7b fragments, according to the manufacturer’s Methyl thiazolyl tetrazolium (MTT) assay for instructions. After washing, membranes were cell proliferation exposed to Kodak XAR-5 film for 20-40 h (Sigma-Aldrich Chemical). A human U6 snRNA Each group cells was seeded at 2 × 103 per well probe (5’-GCAGGGGCCATGCTAATCTTCTCTGTAT- in 96-well plates and cultrued in DMEM supple-

CG-3’) was used as a positive control, with an mented with 10% FBS at 37°C with 5% CO2, exposure time of 15-30 min. until 85% confluent. MTT (Sigma Chemicals) reagent (5 mg/ml) was added to the mainte- cDNA microarray analysis nance cell medium at different time points, and incubated at 37°C for an additional 4 h. The Total RNAs of PCSCs and PCCs were labeled reaction was terminated with 150 μL dimethyl- using Agilent’s Low RNA Input Fluorescent sulfoxide (DMSO, Sigma Chemicals) per well Linear Amplification kit. Cy3-dCTP or Cy5-dCTP and the cells were lysed for 15 min, and the was incorporated during reverse transcription plates were gently shaked per 5 min. of 5 μg total RNAs into cDNA. Different fluores- Absorvance values were determined by using cently labeled cDNA probes were mixed in 30 μl the enzyme linked immunosorbent assay hybridization buffer (3 × SSC, 0.2% SDS, 5 × (ELISA) reader (Model 680, Bio-rad) at 490 nm. Denhardt’s solution and 25% formamide) and applied to the microarray (CapitalBio human Statistical analysis mRNA microarray V2.0, CapitalBio, Beijing, China) following incubation at 42°C for 16 h. Each experiment was performed as least three After hybridization, the slide was washed with times, and data are shown as the mean ± SE 0.2% SDS/2 × SSC at 42°C for 5 min, and then where applicable, and differences were evalu- was washed with 0.2 × SSC at room tempera- ated using Student’s t-tests. The probability of ture for 5 min. The fluorescent images of the P < 0.05 was considered to be statistically hybridized microarray were scanned with an significant. Agilent Whole 4 × 44 microar- ray scanner system (Santa Clara, CA, USA). Results Images and quantitative data of the gene- expression levels were analyzed by Agilent’s CD44+/LIN2B+ PCSCs proliferated more rap- Feature Extraction (FE) software, version 9.5. idly and exhibited multi-drug resistance

Flow cytometric (FCM) analysis of cell cycle by We used a magnetic-activated cell sorting sys- PI staining tem to isolate and enrich the CD44- and LIN28B- overexpressing subpopulation from Each group cells were seeded at 3 × 105 per the primary tumor cells of four human pancre- well in 6-well plates and cultured until 85% con- atic cancer tissues. After isolation, cells were fluent. Each group cells was washed by PBS on quantified by flow cytometry (FCM). CD44+/ three times, then were collected by centrifuga- LIN28B+ PCSCs represented 0.515% ± 0.105% tion (Allegra X-22R, Beckman Coulter) at 1000 of the total population in four primary pancre- g for 5 min. The cell pellets were the resus- atic cancer cells, whereas CD44-/LIN28B- pended in 1 mL of PBS, fixed in 70% ice-cold PCCs represented 91.581% ± 2.961% of the

2646 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Figure 1. Isolation and characterization of CD44+/LIN28B+ cell multi-drug resistance. A. Flow cytometric analysis of the number of pancreatic cells from primary tumor cells from four human pancreatic cancer tissues expressing CD44 and LIN28B. B. Mean numbers of CD44+/LIN28B+ PCSCs and CD44-/LIN28B- PCCs on days 1-6 after pas- sage. *p < 0.05; #p > 0.05 vs. PCCs (n = 3). C. MTT assay results of the effect of cisplatin and gemcitabine hydro- chloride on CD44+/LIN28B+ PCSCs and CD44-/LIN28B-PCCs; **p < 0.01; #p > 0.05 vs. PCCs (n = 3). total population (Figure 1). These results dem- assay to evaluate PCSC and PCC multi-drug onstrated that CD44+/LIN28B+ cells, although resistance. Cisplatin and gemcitabine hydro- very exiguous, could be successfully enriched chloride inhibited both PCSC and PCC growth. using magnetic-activated cell sorting. The However, PCSCs were significantly less suscep- PCSC and PCC proliferation rates were exam- tible to the cytotoxic effects of both drugs ined on days 1-6 after passage. All measure- (Figure 1). Thus, PCSCs were more resistant to ments were repeated in triplicate. There was no cisplatin and gemcitabine hydrochloride than significant difference in the number of cells in PCCs, suggesting that the CD44+/LIN28B+ the two groups on days 0-2 (p > 0.05 vs. PCCs; subpopulation may be resistant to a broad t-tests; n = 3; Figure 1). However, between days spectrum of chemotherapeutics. 3 and 6, PCSCs divided significantly more rap- idly than PCCs (p < 0.05 vs. PCCs; t-tests; n = CD44+/LIN2B+ PCSCs overexpressed stem 3). In addition, the inhibitory rates of cisplatin cell markers (0, 5, 15, 30, and 60 ng/mL) and gemcitabine hydrochloride (0, 10, 20, 50, and 100 ng/mL) We used quantitative reverse transcription- were measured using the MTT proliferation PCR (qRT-PCR) to compare the relative gene

2647 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Figure 2. Increased stem cell marker expression and invasion ability in CD44+/LIN28B+ PCSCs compared to CD44-/LIN28B- PCCs. A. QRT-PCR analysis of the relative expression levels of stem cell marker genes in PCSCs and PCCs; **p < 0.01; #p > 0.05 vs. PCCs (n = 3). B. IF of Nanog, Sox2, Oct4, and LIN28B in PCSCs and PCCs; original magnification × 200. C. Transwell assays of PCSCs and PCCs; **p < 0.01; #p > 0.05 vs. PCCs (n = 3). D. Soft agar colony formation assays of PCSCs and PCCs plated at low density; **p < 0.01; #p > 0.05 vs. PCCs (n = 3). expression levels of several stem cell markers assay, respectively (Figure 2). The Transwell in PCSCs and PCCs; 18S rRNA was used as the assay revealed that significantly fewer PCCs internal control. Nanog, Oct4, Sox2, Tert, invaded compared to PCSCs (invading cell ABCG2, Lin28B, CD44, CD133, and CD117 numbers: PCSCs, 18 ± 2 vs. PCCs, 5 ± 1; p < expression were all significantly higher in 0.01 vs. PCCs; t-tests; n = 3). The soft agar col- PCSCs than in PCCs (Figure 2). Immun- ony formation assay indicated that PCCs ofluorescence staining (IF) confirmed that formed substantially fewer colonies when plat- PCSCs expressed higher levels of the stem cell ed at low density than PCSCs (colony formation markers LIN28B, Oct4, Sox2, and Nanog than efficiency: PCCs, 37.40% ± 3.71% vs. PCSCs, PCCs (p < 0.01 vs. PCCs; t-tests; n = 3; Figure 10.05% ± 1.39%; p < 0.01 vs. PCCs; t-tests; n = 2). These results suggested that the CD44+/ 3). LIN28B+ subpopulation possesses stem cell characteristics. CD44+/LIN2B+ PCSCs induced tumor growth in vivo CD44+/LIN2B+ PCSCs possessed increased migratory and invasive ability To evaluate the tumorigenic capacity of PCSCs and PCCs, 1 × 104 cells were inoculated subcu- The ability of PCSCs and PCCs to migrate and taneously into athymic nude mice. Tumors were invade was determined using the Transwell visible in the PCSC-inoculated mice after 1 migration assay and soft agar colony formation month. However, PCC-inoculated mice did not

2648 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Figure 3. Increased in vivo tumor formation ability in CD44+/LIN28B+ PCSCs compared to CD44-/LIN28B- PCCs. CD44+/LIN28B+ PCSCs or CD44-/LIN28B- PCCs were subcutaneously inoculated into severe combined immuno- deficient (SCID) mice to form in vivo xenografts. A. Representative images of mice with xenograft tumors; yellow rings indicate tumor tissues. B. Tumor growth delay. Tumors formed by the CD44-/LIN28B- PCCs grew more slowly; **P < 0.01, *P < 0.05, and #P > 0.05 vs. PCSC group (n = 4). C. Tumor weight; **P < 0.01 vs. PCSC group (n = 4). D. HE staining of PCSC and PCC tumors revealing cellular heterogeneity of the tumors; original magnification × 200. E. Immunohistochemical staining of the cell proliferation markers Ki67 and Ras indicating weakly positive staining in PCSC tumors and positive or strongly positive staining in PCC tumors; original magnification × 200. exhibit detectable tumors at the same time Ki-67 immunoreactivity (Figure 3). These point (Figure 3). Very small tumors were detect- results were the same as that of Ras protein ed in PCC-injected mice after three months. expression (Figure 3). Representative hema- When the mice were sacrificed four months toxylin and eosin (HE)-stained sections of all after injection, the tumors formed by PCSCs subcutaneous xenograft tumors derived from were significantly heavier than that formed by CD44+/LIN28B+ PCSCs were categorized as PCCs (p < 0.05 vs. PCCs; t-tests; n = 4). The cell moderately or poorly differentiated human pan- proliferation-related protein Ki-67 was ana- creatic carcinoma (Figure 3). Taken together, lyzed in tumor sections using immunohisto- the in vivo xenograft model indicated that low chemistry. The tumors formed by PCSCs dis- numbers of the CD44+/LIN28B+ subpopula- played positive or strongly positive Ki-67 stain- tion have the potential to initiate tumor growth. ing; those formed by PCCs exhibited only weak On the other hand, the results showed that

2649 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Figure 4. Tumor sphere assay and limiting dilution transplantation assays in vitro and in vivo. A. Tumor sphere assay in vitro. The CD44+/LIN28B+ human PCSCs were formatted non-adherent and non-symmetric cell clones in day 3 after plating. Blue arrow indicated the sphere cell clone; the white arrow indicated the single cell. original magnifica- tion × 200. B. The limiting dilution transplantation assays in vivo. All CD44+/LIN28B+ human PCSCs were divided into 5 groups, each group was about 1 × 10 cells (a), 1×102 cells (b), 1 × 103 cells (c), 1 × 104 cells (d), and 1 × 105 cells (e). Each group cells were inoculated s.c in athymic nude mice, respectively. C. HE staining of PCSC tumors revealing cellular heterogeneity of the tumors; original magnification × 200. D. Immunohistochemical staining of the cell proliferation markers Ki67 and Ras indicating positive staining in PCSC tumors; original magnification × 200.

CD44+/LIN28B+ human PCSCs were format- injected mice exhibited detectable tumors at ted sphere cells in day 3 after plating (Figure 4). the same time point (Figure 4). There assays These populations were non-adherent and non- revealed that the quantitative dependence of symmetric cell clones. When spheres were PCSCs oncogenicity in vivo. enzymatically dissociated to single cells, they could give rise to spheres again. This procedure Microarray analysis of cDNA expression pat- could be repeated, and the sphere proliferated terns of CD44+/LIN28B+ PCSCs and CD44-/ faster than the cells under differentiating con- LIN28B- PCCs ditions. Then, the limiting dilution transplanta- tion assay was used to determine the ability To evaluate the difference in gene expression to cause tumor of CD44+/LIN28B+ human patterns of CD44+/LIN28B+ PCSCs and PCSCs. All CD44+/LIN28B+ human PCSCs CD44-/LIN28B- PCCs, we prepared a CapitalBio were divided into 5 groups, each group was human mRNA microarray V2.0 containing about 1 × 10 cells, 1 × 102 cells, 1 × 103 cells, 35000 oligonucleotide probes complementary 1 × 104 cells, and 1 × 105 cells. Each group to known mammalian gene cDNA. Significance cells were inoculated s.c in athymic nude mice, microarray analysis and a fold-change criterion respectively. After 1 month, very small tumors (log10 [PCSCs/PCCs] ratio) of > 2 and q-value < were detected in 1 × 103 cells PCSCs-injected 0.01 were used to identify significant differenc- mice. But, the tumor derived from 1 × 104 cells es. Using these criteria, we identified 1456 PCSCs-injected mice or 1 × 105 cells PCSCs- gene mRNAs that were differentially expressed injected mice was larger significantly than in the PCSCs vs. PCCs [among them, 50 gene above tumor. However, neither 1 × 10 cells mRNAs were upregulated (Table 1); 139 gene PCSCs-injected mice nor 1 × 102 cells PCSCs- mRNAs were downregulated (Table 2)]. Notably,

2650 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Table 1. Upregulated expression genes Genbank Gene Ratio = Log[(PCSCs/ Gene Name Accession Symbol PCCs)] (Ratio > 2) NM_001004317 LIN28B lin-28 homolog B (C. elegans) 3.151903885 NM_002364 MAGEB2 melanoma antigen family B, 2 3.03093709 NM_017410 HOXC13 homeobox C13 2.951947556 NM_000523 HOXD13 homeobox D13 2.757556259 NM_138815 DPPA2 developmental pluripotency associated 2 2.710770922 NM_001008 RPS4Y1 ribosomal protein S4, Y-linked 1 2.702608074 NM_004405 DLX2 distal-less homeobox 2 2.636932885 NM_173553 TRIML2 tripartite motif family-like 2 2.624075551 NM_001008223 C1QL4 complement component 1, q subcomponent-like 4 2.622575612 NM_021796 PLAC1 placenta-specific 1 2.621359161 NM_006237 POU4F1 POU class 4 homeobox 1 2.612859931 NM_033176 NKX2-4 NK2 homeobox 4 2.583539063 NM_031896 CACNG7 calcium channel, voltage-dependent, gamma subunit 7 2.538431113 NM_001039567 RPS4Y2 ribosomal protein S4, Y-linked 2 2.537580073 NM_031959 KRTAP3-2 associated protein 3-2 2.497118083 NM_001004339 ZYG11A zyg-11 homolog A (C. elegans) 2.430087357 NM_006546 IGF2BP1 insulin-like growth factor 2 mRNA binding protein 1 2.427697184 NM_002302 LECT2 leukocyte cell-derived chemotaxin 2 2.378282143 NM_153426 PITX2 paired-like homeodomain 2 2.373912239 NM_024016 HOXB8 homeobox B8 2.370713213 BC041859 LOC780529 uncharacterized LOC780529 2.345185492 NM_012159 FBXL21 F-box and -rich repeat protein 21 (gene/pseudogene) 2.297666083 NM_021192 HOXD11 homeobox D11 2.259208887 NM_015557 CHD5 chromodomain helicase DNA binding protein 5 2.250969281 NM_175868 MAGEA6 melanoma antigen family A, 6 2.239072235 NM_018057 SLC6A15 solute carrier family 6 (neutral transporter), member 15 2.204335394 NM_001112704 VAX1 ventral anterior homeobox 1 2.196077489 NM_001135254 PAX7 paired box 7 2.190715712 NM_005249 FOXG1 forkhead box G1 2.184094348 NM_000853 GSTT1 glutathione S-transferase theta 1 2.176253033 BC039509 LOC643401 uncharacterized LOC643401 2.159535507 NM_025218 ULBP1 UL16 binding protein 1 2.154874271 NM_007129 ZIC2 Zic family member 2 2.118906696 NM_021571 CARD18 caspase recruitment domain family, member 18 2.102380764 NM_003317 NKX2-1 NK2 homeobox 1 2.102375865 NM_016358 IRX4 iroquois homeobox 4 2.101351184 NM_021954 GJA3 gap junction protein, alpha 3, 46kDa 2.098052216 NM_024885 TAF7L TAF7-like RNA polymerase II, TATA box binding protein (TBP)-associated factor, 50kDa 2.094765437 NM_021240 DMRT3 doublesex and mab-3 related transcription factor 3 2.090471549 NM_001122665 DDX3Y DEAD (Asp-Glu-Ala-Asp) box polypeptide 3, Y-linked 2.089953594 NM_001004441 ANKRD34B repeat domain 34B 2.08759533 NM_182767 SLC6A15 solute carrier family 6 (neutral amino acid transporter), member 15 2.086853364 NM_203486 DLL3 delta-like 3 (Drosophila) 2.083388477 NM_152739 HOXA9 homeobox A9 2.057843835 NM_004988 MAGEA1 melanoma antigen family A, 1 (directs expression of antigen MZ2-E) 2.043918381 NM_006361 HOXB13 homeobox B13 2.018953044 NM_018712 ELMOD1 ELMO/CED-12 domain containing 1 2.018650868 NM_003108 SOX11 SRY (sex determining region Y)-box 11 2.011435863 NM_024017 HOXB9 homeobox B9 2.001178387 the scatter plot analysis of the gene mRNA hybridization vs. PCCs mRNA hybridization) indi- microarray results (scatter plot depicts log10- cated differing expression of a certain number transformed ratios obtained from PCSCs mRNA of mRNAs between PCSCs vs. PCCs (Figure 5).

2651 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Table 2. Downregulated expression genes Genbank Gene Ratio = Log[(PCSCs/ Gene Name Accession Symbol PCCs)] (Ratio > 2) NM_001906 CTRB1 chymotrypsinogen B1 4.253352655 NM_001869 CPA2 carboxypeptidase A2 (pancreatic) 4.114128823 NM_001871 CPB1 carboxypeptidase B1 (tissue) 4.073897547 NM_006507 REG1B regenerating islet-derived 1 beta 3.98556258 NM_001007240 GP2 glycoprotein 2 (zymogen granule membrane) 3.957646381 NM_000936 PNLIP pancreatic lipase 3.941809299 NM_001008387 REG3G regenerating islet-derived 3 gamma 3.809492882 NM_006217 SERPINI2 serpin peptidase inhibitor, clade I (pancpin), member 2 3.726041579 NM_007352 CELA3B chymotrypsin-like elastase family, member 3B 3.721707666 NM_005747 CELA3A chymotrypsin-like elastase family, member 3A 3.679697239 NM_019617 GKN1 gastrokine 1 3.650917342 NM_182536 GKN2 gastrokine 2 3.608465143 NM_015849 CELA2B chymotrypsin-like elastase family, member 2B 3.387080468 NM_007352 CELA3B chymotrypsin-like elastase family, member 3B 3.37675489 NM_002909 REG1A regenerating islet-derived 1 alpha 3.339716528 NM_198998 AQP12A aquaporin 12A 3.332356441 NM_033440 CELA2A chymotrypsin-like elastase family, member 2A 3.216462824 NM_001285 CLCA1 chloride channel accessory 1 3.186337576 NM_000928 PLA2G1B phospholipase A2, group IB (pancreas) 3.176664135 NM_007272 CTRC chymotrypsin C (caldecrin) 3.133808678 NM_001443 FABP1 fatty acid binding protein 1, liver 3.10178095 NM_001041 SI sucrase-isomaltase (alpha-glucosidase) 3.077879399 NM_001633 AMBP alpha-1-microglobulin/bikunin precursor 3.043261652 NM_000477 ALB albumin 3.001962782 NM_014276 RBPJL recombination signal binding protein for immunoglobulin kappa J region-like 2.973692081 NM_019010 KRT20 keratin 20 2.961282952 NM_001040462 BTNL8 butyrophilin-like 8 2.925304863 NM_001868 CPA1 carboxypeptidase A1 (pancreatic) 2.908065427 NM_001040105 MUC17 mucin 17, cell surface associated 2.885213886 NM_000349 STAR steroidogenic acute regulatory protein 2.868034449 NM_014471 SPINK4 peptidase inhibitor, Kazal type 4 2.847637265 NM_178161 PTF1A pancreas specific transcription factor, 1a 2.840800222 NM_001145643 PHGR1 /histidine/-rich 1 2.807670721 NM_144696 AXDND1 axonemal light chain domain containing 1 2.794013389 NM_152321 ERP27 endoplasmic reticulum protein 27 2.783485155 NM_001644 APOBEC1 apolipoprotein B mRNA editing enzyme, catalytic polypeptide 1 2.779192438 NM_017625 ITLN1 intelectin 1 (galactofuranose binding) 2.774310879 NM_007272 CTRC chymotrypsin C (caldecrin) 2.756421395

2652 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

NM_005588 MEP1A meprin A, alpha (PABA peptide hydrolase) 2.720645656 NM_005747 CELA3A chymotrypsin-like elastase family, member 3A 2.698310666 NM_003889 NR1I2 nuclear receptor subfamily 1, group I, member 2 2.679493358 NM_032044 REG4 regenerating islet-derived family, member 4 2.675557226 NM_001136485 C11orf86 11 open reading frame 86 2.673767266 NM_001832 CLPS colipase, pancreatic 2.671370467 NM_006229 PNLIPRP1 pancreatic lipase-related protein 1 2.670896325 NM_000492 CFTR cystic fibrosis transmembrane conductance regulator (ATP-binding cassette sub-family C, member 7) 2.663599825 NM_005459 GUCA1C guanylate cyclase activator 1C 2.642911633 NM_003465 CHIT1 chitinase 1 (chitotriosidase) 2.629621307 NM_001076 UGT2B15 UDP glucuronosyltransferase 2 family, polypeptide B15 2.629470167 NM_170741 KCNJ16 potassium inwardly-rectifying channel, subfamily J, member 16 2.592859069 NM_032044 REG4 regenerating islet-derived family, member 4 2.582210072 NM_001907 CTRL chymotrypsin-like 2.578185726 NM_001807 CEL carboxyl ester lipase (bile salt-stimulated lipase) 2.564669478 NM_138938 REG3A regenerating islet-derived 3 alpha 2.540410193 NM_005621 S100A12 S100 calcium binding protein A12 2.507022164 NM_152491 PM20D1 peptidase M20 domain containing 1 2.473104564 NM_001086 AADAC arylacetamide deacetylase (esterase) 2.467135126 NM_003225 TFF1 trefoil factor 1 2.463144163 NM_001003811 TEX11 testis expressed 11 2.450551531 NM_000111 SLC26A3 solute carrier family 26, member 3 2.44439235 NM_003296 CRISP2 -rich secretory protein 2 2.440968817 NM_002443 MSMB microseminoprotein, beta- 2.434490133 NM_005621 S100A12 S100 calcium binding protein A12 2.424419104 NM_000253 MTTP microsomal triglyceride transfer protein 2.423321574 NM_001039112 FER1L6 fer-1-like 6 (C. elegans) 2.406190944 NM_005495 SLC17A4 solute carrier family 17 (sodium phosphate), member 4 2.398680041 NM_001080538 AKR1B15 aldo-keto reductase family 1, member B15 2.396507273 NM_032787 GPR128 G protein-coupled receptor 128 2.381097185 NM_002153 HSD17B2 hydroxysteroid (17-beta) dehydrogenase 2 2.366343469 NM_001185 AZGP1 alpha-2-glycoprotein 1, zinc-binding 2.365710298 NM_006229 PNLIPRP1 pancreatic lipase-related protein 1 2.363652616 NM_016369 CLDN18 claudin 18 2.34702268 NM_001216 CA9 carbonic anhydrase IX 2.344158438 NM_002181 IHH Indian hedgehog 2.334859325 NM_005420 SULT1E1 sulfotransferase family 1E, estrogen-preferring, member 1 2.33310927 NM_170736 KCNJ15 potassium inwardly-rectifying channel, subfamily J, member 15 2.326162651 NM_020299 AKR1B10 aldo-keto reductase family 1, member B10 (aldose reductase) 2.3217617 NM_020299 AKR1B10 aldo-keto reductase family 1, member B10 (aldose reductase) 2.321622661 NM_021969 NR0B2 nuclear receptor subfamily 0, group B, member 2 2.295677518

2653 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

NM_005423 TFF2 trefoil factor 2 2.283895997 NM_001101404 SH2D7 SH2 domain containing 7 2.275404367 NM_019596 C21orf62 open reading frame 62 2.275342284 NM_004212 SLC28A2 solute carrier family 28 (sodium-coupled nucleoside transporter), member 2 2.271658804 NM_005814 GPA33 glycoprotein A33 (transmembrane) 2.269666309 NM_005950 MT1G metallothionein 1G 2.263322944 NM_033050 SUCNR1 succinate receptor 1 2.258936505 NM_001074 UGT2B7 UDP glucuronosyltransferase 2 family, polypeptide B7 2.257942574 NM_014080 DUOX2 dual oxidase 2 2.251638427 NM_007193 ANXA10 annexin A10 2.24536539 NM_080870 DPCR1 diffuse panbronchiolitis critical region 1 2.227683528 NM_153343 ENPP6 ectonucleotide pyrophosphatase/phosphodiesterase 6 2.212974443 NM_001080468 SYCN syncollin 2.205592389 NM_002927 RGS13 regulator of G-protein signaling 13 2.204853321 NM_198477 CXCL17 chemokine (C-X-C motif) ligand 17 2.202850396 NM_001080405 CEACAM18 carcinoembryonic antigen-related cell adhesion molecule 18 2.20198379 NM_001721 BMX BMX non-receptor tyrosine kinase 2.200051762 NM_000384 APOB apolipoprotein B (including Ag(x) antigen) 2.199449748 NM_005951 MT1H metallothionein 1H 2.196296242 NM_000128 F11 coagulation factor XI 2.195335353 NM_001073 UGT2B11 UDP glucuronosyltransferase 2 family, polypeptide B11 2.194853528 NM_003963 TM4SF5 transmembrane 4 L six family member 5 2.194043134 NM_000341 SLC3A1 solute carrier family 3 (cystine, dibasic and neutral amino acid transporters, activator of cystine, dibasic and neutral amino acid transport), member 1 2.19054256 NM_007072 HHLA2 HERV-H LTR-associating 2 2.186569068 NM_001169 AQP8 aquaporin 8 2.185414667 NM_006061 CRISP3 cysteine-rich secretory protein 3 2.174405317 NM_004063 CDH17 cadherin 17, LI cadherin (liver-intestine) 2.169387832 NM_024743 UGT2A3 UDP glucuronosyltransferase 2 family, polypeptide A3 2.163743137 NM_170736 KCNJ15 potassium inwardly-rectifying channel, subfamily J, member 15 2.156444411 NM_021161 KCNK10 potassium channel, subfamily K, member 10 2.135896894 NM_001080527 MYO7B VIIB 2.135028806 NM_138969 SDR16C5 short chain dehydrogenase/reductase family 16C, member 5 2.117283297 NM_203395 IYD iodotyrosine deiodinase 2.109420053 NM_194439 RNF212 ring finger protein 212 2.108893597 NM_001085382 PSAPL1 prosaposin-like 1 (gene/pseudogene) 2.106102812 NM_004132 HABP2 hyaluronan binding protein 2 2.105420932 NM_001807 CEL carboxyl ester lipase (bile salt-stimulated lipase) 2.104908892 NM_001040105 MUC17 mucin 17, cell surface associated 2.103367532 NM_003015 SFRP5 secreted frizzled-related protein 5 2.102989014 NM_006418 OLFM4 olfactomedin 4 2.092198248 NM_004963 GUCY2C guanylate cyclase 2C (heat stable enterotoxin receptor) 2.087387429

2654 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

NM_001010903 C6orf222 chromosome 6 open reading frame 222 2.083954408 NM_145650 MUC15 mucin 15, cell surface associated 2.077301731 NM_002639 SERPINB5 serpin peptidase inhibitor, clade B (ovalbumin), member 5 2.076055664 NM_000277 PAH hydroxylase 2.075090795 NM_014465 SULT1B1 sulfotransferase family, cytosolic, 1B, member 1 2.068579425 NM_206819 MYBPC1 myosin binding protein C, slow type 2.057838856 NM_001079807 PGA3 pepsinogen 3, group I (pepsinogen A) 2.047264536 NM_207581 DUOXA2 dual oxidase maturation factor 2 2.042660099 NM_080658 ACY3 aspartoacylase (aminocyclase) 3 2.037460662 NM_152311 CLRN3 clarin 3 2.036559534 NM_001010857 LELP1 late cornified envelope-like proline-rich 1 2.033215345 NM_182983 HPN hepsin 2.028306399 NM_024533 CHST5 carbohydrate (N-acetylglucosamine 6-O) sulfotransferase 5 2.027032681 NM_005379 MYO1A myosin IA 2.024188703 NM_024921 POF1B premature ovarian failure, 1B 2.020794379 NM_212557 AMTN amelotin 2.017073209 NM_003226 TFF3 trefoil factor 3 (intestinal) 2.013612065 NM_000035 ALDOB aldolase B, fructose-bisphosphate 2.011105368

2655 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

Figure 5. Microarray analysis of cDNA expression patterns in CD44+/LIN28B+ PCSCs vs. CD44-/LIN28B- PCCs. A. Cluster analysis of differentially expressed gene mRNAs in CD44+/LIN28B+ PCSCs vs. CD44-/LIN28B- PCCs. B. Scatter plot indicating differing mRNA expression between CD44+/LIN28B+ PCSCs vs. CD44-/LIN28B- PCCs.

Figure 6. Suppression of Lin28b expression in CD44+/LIN2B+ PCSCs decreased proliferation while stimulating let- 7b expression to interfere with CCND1 expression. A. Western blotting confirming that CCND1 expression was sig- nificantly decreased in siRNA-Lin28b–transfected PCSCs compared to siRNA-Mock-transfected PCSCs. Glyceralde- hyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control; **P < 0.01, *P < 0.05, and #P > 0.05 vs. siRNA-Mock (n = 3). B. Northern blotting indicating a strong let-7b hybridization signal in siRNA-Lin28b-transfected PCSCs compared to siRNA-Mock-transfected PCSCs. C. MTT assay of the effect of proliferation inhibition in siRNA- Lin28b-transfected PCSCs compared to siRNA-Mock-transfected PCSCs; **p < 0.01, #p > 0.05 vs. siRNA-Mock; (n = 3). D. FCM showing that compared to siRNA-Mock-transfected cells, siRNA-Lin28b-transfected PCSCs were arrested in the G0/G1 phase and the percentage of S-phase cells was significantly decreased.

2656 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

These results demonstrate that gene mRNA that they were closely associated with both expression patterns in PCSCs were significantly high proliferation and invasion ability, and diffi- different from that of PCCs. cult early-stage diagnosis and poor recurrence- free prognosis. However, knowledge of the Suppression of endogenous Lin28b expres- biological characteristics of PCSCs is limited. sion in CD44+/LIN2B+ PCSC decreased prolif- Although it has been reported that the CD31+/ eration while stimulating let-7b expression to CD45+, Hoechst 33342-/CD133+/ALDH1+, interfere with CCND1 expression ESA+/CD44+, and CD24+/CD44+ subpopula- tions in pancreatic cancer not only overexpress To determine whether endogenous Lin28b stem cell markers [1, 3, 4, 6, 12], but also interfered with expression of the promoter exhibit high self-renewal and migratory ability miRNA let-7b and influenced CD44+/LIN2B+ and multi-drug resistance, we believe that PCSC proliferation, siRNA-Lin28b and siRNA- there are other subpopulations that possess Mock were transfected into CD44+/LIN2B+ PCSC characteristics in primary pancreatic PCSCs. The efficiency of the transfected siRNA cancer tissues. Referring to previous studies, on mRNA and protein expression was detected we found that LIN28B regulated not only embry- by western and northern blotting, respectively. onic stem cells, but also cancer cell self-renew- Western blotting (Figure 6) showed that LIN28B al. When exogenous Lin28 was overexpressed and CCND1 expression in PCSCs transfected in host cells, the proliferation ability in these with siRNA-Mock (0.700 ± 0.078; 0.897 ± cells increased significantly. This indicates that 0.041, respectively) was higher than that in LIN28B is a positive regulator of cell prolifera- PCSCs transfected with siRNA-Lin28b (0.170 ± tion. As CD44 expression occurs in a wide vari- 0.031; 0.433 ± 0.027, respectively; p < 0.05 ety of CSCs, we used CD44 and LIN28B as CSC vs. siRNA-Mock; t-tests; n = 3). These results markers to sort PCSCs from primary pancreatic demonstrated that siRNA-Lin28b specifically interfered with LIN28B expression in PCSCs; at cancer cells. Although the CD44+/LIN28B+ the same time, it weakened CCND1 expression. subpopulation share in the overall cell popula- Northern blotting revealed a strong let-7b tion was very low, it exists at a certain ratio in hybridization signal in PCSCs transfected with the tumor tissues of many pancreatic cancer siRNA-Lin28b compared with PCSCs transfect- patients. We determined the CSC characteris- ed with siRNA-Mock (Figure 5). To determine tics of CD44+/LIN28B+ and CD44-/LIN28B- whether LIN28B interference would suppress pancreatic cancer cells. We found that not only PCSC proliferation, inhibition at 0, 24, 48, and self-renewal ability, but also migratory ability 72 h was measured by MTT assay (Figure 6). At and multi-drug resistance were higher in 48 h and 72 h, PCSCs transfected with siRNA- CD44+/LIN28B+ cells than in CD44-/LIN28B- Mock were significantly less susceptible to the cells. Moreover, there was high expression of proliferation inhibitory effect than PCSCs trans- stem cell markers by CD44+/LIN28B+ cells, fected with siRNA-Lin28b (p < 0.01 vs. siRNA- and the cells exhibited high tumorigenic ability Mock; t-tests; n = 3; Figure 6). FCM demon- in vivo. In view of these results, we believe strated significant cell cycle arrest of PCSCs that the CD44+/LIN28B+ subpopulation in transfected with siRNA-Lin28b. Compared with pancreatic cancer cells also possesses PCSC PCSCs transfected with siRNA-Mock, PCSCs characteristics. transfected with siRNA-Lin28b were arrested in the G0/G1 phase and the percentage of The other reason for using LIN28B as a PCSC S-phase cells was significantly decreased p( < sorting marker is that it regulates Oct4 expres- 0.05 vs. siRNA-Mock; t-tests; n = 3; Figure 6). sion in CSCs. Qiu and Huang [26] reported that These data indicate that not only was endoge- Lin28 mediated post-transcriptional regulation nous let-7b expression promoted and CCND1 of Oct4 expression in human embryonic stem inhibited in the CD44+/LIN28B+ PCSCs, but cells. They found that Lin28 binds Oct4 mRNA also that proliferation of the subpopulation directly through high-affinity sites within its was weakened when endogenous LIN28B was coding region and that interaction between suppressed. Lin28 and RNA helicase A may play a part in the observed regulation. They further demon- Discussion strated that decreasing RNA helicase A levels impaired Lin28-dependent stimulation of Since pancreatic CSCs were first found in pan- translation in a reporter system [26]. In view of creatic cancer tissue, it was generally believed many reports demonstrating that CSCs also

2657 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion overexpress stem cell markers, such as Oct4, [4] Han JB, Sang F, Chang JJ, Hua YQ, Shi WD, Sox2, and Nanog, we hypothesized that pan- Tang LH, Liu LM. Arsenic trioxide inhibits viabil- creatic cancer cells expressed high levels of ity of pancreatic cancer stem cells in culture stem cell markers, such as the LIN28B- and in a xenograft model via binding to SHH- Gli. Onco Targets Ther 2013; 6: 1129-1138. overexpressing subpopulation. Our findings [5] Lu Y, Zhu H, Shan H, Lu J, Chang X, Li X, Lu J, demonstrated that stem cell marker expres- Fan X, Zhu S, Wang Y, Guo Q, Wang L, Huang Y, sion in CD44+/LIN28B+ PCSCs was significant- Zhu M, Wang Z. Knockdown of Oct4 and Nanog ly higher than that in CD44-/LIN28B- cells. expression inhibits the stemness of pancreatic cancer cells. Cancer Lett 2013; 340: 113-123. In conclusion, CD44+/LIN28B+ PCSCs not only [6] Li C, Heidt DG, Dalerba P, Burant CF, Zhang L, express high levels of stem cell markers, but Adsay V, Wicha M, Clarke MF, Simeone DM. also possess strong self-renewal and migratory Identification of pancreatic cancer stem cells. ability and multi-drug resistance in vitro and Cancer Res 2007; 67: 1030-1037. are tumorigenic in vivo. [7] Liu T, Xu F, Du X, Lai D, Liu T, Zhao Y, Huang Q, Jiang L, Huang W, Cheng W, Liu Z. Establishment Acknowledgements and characterization of multi-drug resistant, prostate carcinoma-initiating stem-like cells from human cell lines 22RV1. This work was supported by grant from National Mol Cell Biochem 2010; 340: 265-273. Natural Science Foundation of China (No. [8] Ponti D, Costa A, Zaffaroni N, Pratesi G, 81202811), and Project funded by China Petrangolini G, Coradini D, Pilotti S, Pierotti Postdoctoral Science Foundation (No. 2014- MA, Daidone MG. Isolation and in vitro propa- M550250), and Shanghai Municipal Health gation of tumorigenic breast cancer cells with Bureau Fund (No. 20124320) to Te Liu. stem/progenitor cell properties. Cancer Res 2005; 65: 5506-5511. Disclosure of conflict of interest [9] Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. None. Nature 2001; 414: 105-111. [10] Marx J. Cancer research. Mutant stem cells Address correspondence to: Drs. Dayong Jin and may seed cancer. Science 2003; 301: 1308- Weiqi Lu, Department of General Surgery, Zhong- 1310. [11] Pardal R, Clarke MF, Morrison SJ. Applying the shan Hospital, Fudan University, Shanghai 200032, principles of stem-cell biology to cancer. Nat China. Tel: 86-21-64041990; Fax: 86-21- Rev Cancer 2003; 3: 895-902. 64038038; E-mail: [email protected]; Te Liu, [12] Van den Broeck A, Vankelecom H, Van Delm Shanghai Geriatric Institute of Chinese Medicine, W, Gremeaux L, Wouters J, Allemeersch J, Longhua Hospital, Shanghai University of Traditional Govaere O, Roskams T, Topal B. Human pan- Chinese Medicine, Building C, 365 Xiangyang Road, creatic cancer contains a side population ex- Shanghai 200031, China. Tel: 86-21-64720010; pressing cancer stem cell-associated and Fax: 86-21-64720010; E-mail: [email protected] prognostic genes. PLoS One 2013; 8: e73968. [13] Bapat SA, Mali AM, Koppikar CB, Kurrey NK. References Stem and progenitor-like cells contribute to the aggressive behavior of human epithelial ovari- [1] Wang X, Liu Q, Hou B, Zhang W, Yan M, Jia H, Li an cancer. Cancer Res 2005; 65: 3025-3029. H, Yan D, Zheng F, Ding W, Yi C, Hai W. [14] Park DM, Rich JN. Biology of glioma cancer Concomitant targeting of multiple key tran- stem cells. Mol Cells 2009; 28: 7-12. scription factors effectively disrupts cancer [15] Lei XX, Xu J, Ma W, Qiao C, Newman MA, stem cells enriched in side population of hu- Hammond SM, Huang Y. Determinants of man pancreatic cancer cells. PLoS One 2013; mRNA recognition and translation regulation 8: e73942. by Lin28. Nucleic Acids Res 2012; 40: 3574- [2] Siegel R, Naishadham D, Jemal A. Cancer sta- 3584. tistics, 2012. CA Cancer J Clin 2012; 62: 10- [16] Moss EG, Tang L. Conservation of the heter- 29. ochronic regulator Lin-28, its developmental [3] Herreros-Villanueva M, Zhang JS, Koenig A, expression and microRNA complementary Abel EV, Smyrk TC, Bamlet WR, de Narvajas AA, sites. Dev Biol 2003; 258: 432-442. Gomez TS, Simeone DM, Bujanda L, Billadeau [17] Peng S, Chen LL, Lei XX, Yang L, Lin H, DD. SOX2 promotes dedifferentiation and im- Carmichael GG, Huang Y. Genome-wide stud- parts stem cell-like features to pancreatic can- ies reveal that Lin28 enhances the translation cer cells. Oncogenesis 2013; 2: e61. of genes important for growth and survival of

2658 Am J Cancer Res 2015;5(9):2643-2659 LIN28B promotion of PCSC proliferation and invasion

human embryonic stem cells. Stem Cells [23] Xu B, Zhang K, Huang Y. Lin28 modulates cell 2011; 29: 496-504. growth and associates with a subset of cell cy- [18] Xu F, Wang H, Zhang X, Liu T, Liu Z. Cell prolif- cle regulator mRNAs in mouse embryonic stem eration and invasion ability of human chorio- cells. RNA 2009; 15: 357-361. carcinoma cells lessened due to inhibition of [24] Schultz J, Lorenz P, Gross G, Ibrahim S, Kunz Sox2 expression by microRNA-145. Exp Ther M. MicroRNA let-7b targets important cell cycle Med 2013; 5: 77-84. molecules in malignant melanoma cells and [19] Qin W, Ren Q, Liu T, Huang Y, Wang J. interferes with anchorage-independent growth. MicroRNA-155 is a novel suppressor of ovarian Cell Res 2008; 18: 549-557. cancer-initiating cells that targets CLDN1. [25] Yu F, Yao H, Zhu P, Zhang X, Pan Q, Gong C, FEBS Lett 2013; 587: 1434-1439. Huang Y, Hu X, Su F, Lieberman J, Song E. let-7 [20] Liu T, Huang Y, Liu J, Zhao Y, Jiang L, Huang Q, regulates self renewal and tumorigenicity of Cheng W, Guo L. MicroRNA-122 influences the breast cancer cells. Cell 2007; 131: 1109- development of sperm abnormalities from hu- 1123. man induced pluripotent stem cells by regulat- [26] Qiu C, Ma Y, Wang J, Peng S, Huang Y. Lin28- ing TNP2 expression. Stem Cells Dev 2013; mediated post-transcriptional regulation of 22: 1839-1850. Oct4 expression in human embryonic stem [21] Cheng W, Liu T, Wan X, Gao Y, Wang H. cells. Nucleic Acids Res 2010; 38: 1240-1248. MicroRNA-199a targets CD44 to suppress the tumorigenicity and multidrug resistance of -initiating cells. FEBS J 2012; 279: 2047-2059. [22] Zhang L, Liu T, Huang Y, Liu J. microRNA-182 inhibits the proliferation and invasion of hu- man lung adenocarcinoma cells through its ef- fect on human cortical -associated pro- tein. Int J Mol Med 2011; 28: 381-388.

2659 Am J Cancer Res 2015;5(9):2643-2659